1. Problem Formulation
When driving a bike or a car, resources used to interact with a mobile device (phone) are limited. Operator's eyes are occupied most of the time for monitoring the traffic and hands are used for other tasks such as operating vehicle steering. However, current phone interaction is highly dependent on a visual feedback from a display, and precise motor operations. More simple and robust techniques are needed for the most important tasks, such as, e.g., incoming call handling.
2. Prior Art Solutions
Speech commands have been used in vehicles. However, when driving (especially a bike), surrounding noises reduce a reliability of an automatic speech recognition. Using the speech recognition alone is also difficult; the reliability of the recognition is much higher when an activation button is used to explicitly start the recognition. This limits the usefulness of the speech recognition, since instead of pressing a “start recognition” key, the user can as easily press an “answer incoming call” key.
Some cars also have dedicated keys installed for simple phone commands (e.g., Send/End keys). This solution is pretty good in terms of usability and safety, but requires complex installation.
Acceleration sensors can be used to detect simple tap or tilt gestures. For example, the user can simply tap the phone front to answer a call. However, the acceleration-sensing is sensitive to external disturbances, especially when driving on a bumpy road, or when using a non-spring-supported vehicle (e.g. most bikes do not have springs).